1. Field of the Invention
The present general inventive concept relates to a power supply device, a micro server having the same, and a power supply method. More particularly, the present general inventive relates to a power supply device that can operate selectively a plurality of power supplies depending on load, a micro server having the same, and a power supply method.
2. Description of the Related Art
Due to recent advancements in high-speed Internet and intranet technologies, server technology that allows a large amount of data to be processed at high speed has been developed. Accordingly, rack-mount cluster server technology has emerged, but has problems due to its bulkiness and significantly high amount of power consumption. Furthermore, since server modules are connected with each other by cables in a rack-mount cluster setting, system expansion is limited.
To solve the above problems, a micro server that uses a processor module has been developed. The processor module refers to a thin, modular expansion server that is inserted and operated in a main body of a micro server system, and does not include stacking rack servers horizontally like the rack-mount servers. The micro server is referred to as a high-density server in the sense that a large number of servers can be inserted and installed in a narrow space. The micro server contains key elements of the server such as at least one central processing unit (CPU), a memory device, an operating system, etc., and performs functions as a server by using power, input and output, incidental devices, and various control functions supported by the main body.
In order to ensure reliable and smooth operation of the server, the power supply device of the server has been designed to be redundant to compensate for any potential power supply failure.
However, in the sense that the power supply device of the server is designed to operate at the highest level of efficiency in a load of a specific region due to primary/secondary transformer design characteristics, a plurality of power supplies configured as the redundant power supply device supplies power to the server system at an equal rate. As a result, the efficiency of the server system is less than 50% in a low load condition.
Moreover, since the efficiency of a power supply unit (PSU) is lessened, a main power supply device eventually results in a power loss of the whole system. Accordingly, efficient design and operation of the power supply device implementing a plurality of PSUs in parallel is required. However, since the overall system loads are supplied in parallel, the PSUs are driven in the undesired low load condition such that a large amount of loss is generated.
In particular, since loss caused by power delivery occupies a large proportion of the power management in data centers, servers, etc., there is a need for a high-efficiency design of the power supply.